John Lewis has unveiled a groundbreaking AI tool to verify the age of customers purchasing knives online, marking a shift in how retailers address legal requirements for the sale of bladed items.

According to the AI provider, their AI (which estimates the age of the user from their image) is “Better Than Human Judgement”.

Why Is John Lewis Introducing AI for Age Verification?

The decision to implement an AI-driven facial age estimation system stems from a broader effort to prevent underage access to knives amidst increasing scrutiny of age verification processes. This move comes as part of the retailer’s commitment to safety and compliance with government regulations while reintroducing online knife sales after a 15-year hiatus, against the backdrop of high-profile cases, such as the tragic murders linked to underage perpetrators purchasing knives online, which have reignited debates about stricter controls on bladed items.

John Lewis stopped selling knives online in 2009 due to the difficulty of verifying buyers’ ages effectively. By 2022, the retailer went a step further, removing cutlery knives from its online catalogue. However, the retailer has now reintroduced these products, citing confidence in the efficacy of AI-powered age estimation technology to meet strict legal and ethical requirements.

As a spokesperson for John Lewis recently explained: “We take safety incredibly seriously, and in line with strict government guidelines, have added an additional layer of security when customers purchase knives online.”

How Does the AI Tool Work?

The facial age estimation technology, developed by British company Yoti, analyses a photograph of the customer’s face to determine whether they are over 18. This streamlined process occurs at the point of purchase and takes only a few seconds. Customers are prompted to enable their device’s camera and position their face within a frame on the screen, akin to using a passport photo booth.

The AI system then estimates the individual’s age and immediately deletes the image once verification is complete. If the system determines the customer is over 18, they can proceed to checkout. For those who do not pass this initial check, an alternative verification method is available, allowing customers to upload a photo of their ID and a selfie to confirm their identity. Accepted forms of ID include passports, driving licences, and other official identification cards.

In addition to this online verification, a second layer of age checking occurs at delivery. For example, Royal Mail or DPD couriers require customers to present valid photo identification, such as a passport or driving licence, before handing over the parcel. If the recipient cannot provide proof of age, the item is returned to John Lewis, and a refund is issued.

What Technology Powers the Tool?

Yoti’s AI age estimation system relies on advanced machine learning algorithms trained on millions of images paired with verified ages. The technology does not rely on facial recognition, meaning it does not match the scanned face to a database of images or identities. Instead, it estimates age based on facial characteristics and deletes the image immediately after processing.

Better Than Human Judgement, Says Yoti

Yoti claims the system offers superior accuracy compared to human judgment. For example, for individuals aged 13–24, the tool estimates age within a margin of 1.3 years. The tool’s accuracy rate for correctly identifying 13–17-year-olds as under 18 is an impressive 99.3 per cent, with negligible variance across different skin tones, according to a 2023 white paper. The system also incorporates anti-spoofing technology to prevent attempts to bypass the check using photos, masks, or deepfake videos.

The Benefits of the System

The reintroduction of online knife sales by John Lewis demonstrates the potential of AI to address regulatory challenges while improving customer convenience. For the retailer, the technology enables compliance with laws requiring age verification at the point of sale and delivery.

The integration of this technology is expected to reduce the administrative burden associated with manual ID checks while offering customers a seamless and fast checkout process. Also, the system helps protect public safety by reducing the risk of knives falling into the hands of minors.

Commander Stephen Clayman of the National Police Chiefs’ Council was recently quoted (in The Times) praising the initiative, saying: “We welcome technology which can help to ensure knives do not end up in the wrong hands. Responsible retailing is a key element in this, and innovations like this are a step in the right direction.”

Privacy-Focused

One other key compliance benefit of the tool is that it’s also privacy-focused, as no images or personal data are stored, shared, or used for further training. This ensures compliance with data protection regulations and alleviates concerns about surveillance.

Challenges and Criticisms

Despite its benefits, the system is not without its challenges. One concern is the tool’s reliance on accurate camera functionality, which may exclude customers who lack access to modern devices or are unfamiliar with using such technology. Customers experiencing technical issues may find the process cumbersome, particularly if they need to switch to the manual ID verification method.

Another issue lies in potential inaccuracies. While the system boasts a high degree of accuracy, its effectiveness diminishes slightly for edge cases, e.g. individuals who appear significantly older or younger than their actual age. Critics have also pointed out that, although rare, the slight variation in accuracy across skin tones highlights an area where further refinement is needed.

Also, broader societal concerns remain about over-reliance on AI in public-facing applications. Privacy advocates, for example, have cautioned against the widespread adoption of AI for age verification, arguing that such systems, while anonymised, may normalise invasive technologies.

A Retail Trend?

It should be noted here that John Lewis is not alone in adopting AI for age verification. For example, Yoti’s technology is already used by social media platforms, alcohol retailers, and other businesses requiring age-restricted transactions. The wider adoption of AI age estimation tools could represent a turning point in retail, enabling businesses to meet regulatory demands while enhancing customer experience.

With the UK government considering stricter regulations on knife sales, including potential requirements for multiple forms of ID, John Lewis’ proactive use of technology may set a precedent for other retailers. As the national conversation around knife crime continues, innovations like this highlight the role of technology in tackling complex societal challenges.

By blending cutting-edge AI with robust checks and balances, John Lewis may have found a way to navigate a path forward in a contentious area of retail, but the journey is far from over. How other retailers respond, and whether customers embrace or resist this technological shift, remains to be seen.

What Does This Mean For Your Business?

By integrating advanced facial age estimation technology into its operations, the retailer has taken a proactive, technology-led approach to tackling what has been, up until now, a complex issue. This initiative has allowed John Lewis to re-enter the online knife market after years of hiatus, balancing customer convenience with security and showcasing the transformative potential of AI in retail.

However, as with any technological innovation, the implementation of such systems raises broader questions. While the facial age estimation tool offers a streamlined and privacy-focused solution, it is not without limitations. Issues such as accessibility for those without modern devices, potential inaccuracies at the margins of the system’s age-detection capabilities, and ongoing concerns about the normalisation of AI in public-facing applications highlight areas for further development and debate.

The integration of a secondary verification step, requiring proof of age upon delivery, ensures an additional layer of security. This dual-layered system strengthens compliance and demonstrates John Lewis’ commitment to responsible retailing. At the same time, it underscores the importance of redundancy in technological systems to account for potential failures or inaccuracies in AI processes.

While this initiative could position John Lewis as a leader in leveraging AI for compliance, it may also signal the beginning of a broader trend within the retail sector. As more businesses explore AI-based solutions for age-restricted sales, a wider conversation about the ethical, practical, and societal implications of these technologies is inevitable. The delicate balance between leveraging innovation for efficiency and ensuring equitable access and fairness will be crucial for widespread acceptance.

John Lewis’ adoption of AI age verification could offer a glimpse into the future of retail. It demonstrates how technology can address pressing regulatory and societal challenges, albeit with some caveats. Whether this approach becomes an industry standard or prompts further refinements in the application of AI remains to be seen, but what is clear is that this marks an important moment in the ongoing evolution of responsible retail practices. For now, John Lewis can say it has set a benchmark, but the effectiveness and reception of this technology will ultimately shape its long-term role in retail. No doubt other retailers will be watching with interest.

LinkedIn, the professional networking giant owned by Microsoft, is under fire as a new lawsuit alleges the platform disclosed the private messages of its Premium customers to train generative AI models without consent.

The lawsuit, filed in California on behalf of Alessandro De La Torre and millions of other Premium subscribers, accuses LinkedIn of breaching contractual promises and violating US privacy laws.

The controversy centres on LinkedIn’s policy changes in 2024, which allowed user data to be used for AI training purposes. While LinkedIn exempted users in countries with stringent privacy regulations (e.g. the UK, EU, and Canada) from this practice, US users were automatically enrolled in the data-sharing programme unless they manually opted out. Crucially, the lawsuit alleges that LinkedIn extended this data-sharing to include the contents of private InMail messages, which often contain sensitive personal and professional information.

The lawsuit highlights the potential implications for users, stating that these private messages could include “life-altering information about employment, intellectual property, compensation, and other personal matters.” This, the plaintiff argues, breaches the LinkedIn Subscription Agreement (LSA), which explicitly assures Premium customers that their confidential information will not be disclosed to third parties. The complaint also points out that LinkedIn’s alleged failure to notify customers of these changes undermines user trust and constitutes a breach of the US Stored Communications Act.

LinkedIn has denied the allegations, labelling them as “false claims with no merit.” However, for many, the platform’s response to the privacy concerns raised last year casts a shadow over its denials. For example, in August 2024, LinkedIn introduced a setting allowing users to opt out of data-sharing for AI training, but this was enabled by default, raising questions about informed consent. Also, the platform discreetly updated its privacy policy in September 2024 to include the use of user data for AI training, with a notable caveat: opting out would not affect data already used to train models.

Some legal commentators have noted that this case could set a significant precedent for how social media platforms and tech companies handle user data in the age of AI. For example, as the plaintiff’s attorney, Rafey Balabanian, says: “This lawsuit underscores a growing tension between innovation and privacy,” and that “LinkedIn’s actions, if proven, represent a serious breach of trust, particularly given the sensitive nature of the information involved.”

The potential fallout for LinkedIn could extend beyond the courtroom. Premium customers, who pay up to $169.99 per month for features like InMail messaging and enhanced privacy, may, for example, choose to reconsider their subscriptions if these allegations prove true. Also, the case draws attention to the broader issue of how companies disclose and manage data for AI development, a concern that has already prompted regulatory scrutiny in regions like the UK and EU. Notably, the UK Information Commissioner’s Office (ICO) had earlier pressed LinkedIn to halt the use of UK user data for AI training, to which LinkedIn had agreed.

For users, this lawsuit serves as a reminder of the need to scrutinise privacy settings and policies. If successful, the plaintiffs seek damages, statutory penalties of $1,000 per affected user, and the deletion of any AI models trained using their data. With LinkedIn potentially facing financial and reputational damage, this case could act as a catalyst for greater transparency and accountability in the tech industry. Whether LinkedIn’s alleged actions were an oversight or a deliberate strategy to accelerate AI innovation, the coming months will undoubtedly shape the future of user privacy in the digital age.

Cloudflare’s latest DDoS Threat Report for Q4 2024 highlights a dramatic surge in Distributed Denial of Service (DDoS) attacks, including a record-breaking 5.6 Tbps assault.

The web security and infrastructure company’s report reveals a 53 per cent year-over-year rise in DDoS activity, with Cloudflare blocking 21.3 million attacks in 2024, 6.9 million of which occurred in Q4, a staggering 83 per cent increase from the same period in 2023!

The largest attack, a 5.6 Tbps assault by a Mirai-variant botnet of over 13,000 IoT devices, targeted an ISP in Eastern Asia. Cloudflare says it mitigated it autonomously within seconds, preventing any disruption. Hyper-volumetric attacks exceeding 1 Tbps grew by 1,885 per cent quarter-over-quarter, reflecting the increasing scale and intensity of these threats. Nearly half of all attacks targeted OSI Layers 3 and 4, with the remainder focused on HTTP-based attacks, predominantly launched by botnets exploiting IoT devices.

Cloudflare’s report also highlighted how emerging attack methods like Memcached and BitTorrent DDoS vectors have seen dramatic growth, and ransom-motivated attacks surged by 78 per cent compared to Q3. The report also identifies telecommunications and marketing as the most attacked industries, with China, the Philippines, and Taiwan being key hotspots. Cloudflare says those responsible for the attacks include competitors, state-sponsored groups, and disgruntled users, highlighting diverse motives behind these incidents.

To counter these growing threats, businesses should deploy always-on, automated DDoS protection, secure all connected devices, and adopt proactive defence strategies. With attacks becoming faster and more sophisticated, real-time mitigation and robust security are critical to minimising risk.

Scientists at the Massachusetts Institute of Technology (MIT) have unveiled a new generation of tiny insect-inspired flying robots that could revolutionise agriculture by offering a mechanical alternative to natural pollinators.

The Vision Behind the Robotic Pollinators

Pollination is one of the most critical processes in food production, yet the decline in bee populations due to habitat loss, pesticides, and also with climate change posing a growing threat to global agriculture. Enter the robotic insect, a tiny flying marvel designed to fill the gap left by natural pollinators. Developed by a team led by Associate Professor Kevin Chen, head of MIT’s Soft and Micro Robotics Laboratory, these robots could “swarm out of mechanical hives” to pollinate plants with precision.

“With the improved lifespan and precision of this robot, we are getting closer to some very exciting applications, like assisted pollination,” Chen explains. His team’s latest innovation, showcased in Science Robotics, represents a significant leap forward in terms of flight performance and potential practical applications.

What Are These Robots?

The robots, weighing less than a paperclip, are designed to mimic the flight patterns of insects such as bees. Each robot features four units equipped with flapping wings powered by artificial muscles. These soft actuators are constructed from layers of elastomer (a flexible, rubber-like material that can stretch and return to its original shape) sandwiched between carbon nanotube electrodes, allowing the wings to beat at high frequencies.

What sets this newest version apart from earlier efforts is its durability and efficiency. For example, the previous models could only fly for about 10 seconds before succumbing to mechanical strain, whereas the revamped version can hover for over 1,000 seconds (nearly 17 minutes) without degrading its performance. This remarkable improvement stems from a complete overhaul of the robot’s wing and transmission design.

A New Standard in Robotic Agility and Precision

The latest version of the robot bug is not just durable, it’s also highly agile. For example, it can perform complex acrobatic manoeuvres, such as double aerial flips and body rolls, and trace remarkably specific flight paths with incredible precision. The scientists have even been able to make a swarm of the robot bugs spell out “M-I-T” mid-flight (rather like drone displays). These capabilities are underpinned by advanced control systems and a redesigned wing structure that reduces mechanical stress.

For example, as explained by Chen: “Compared to the old robot, we can now generate control torque three times larger than before, which is why we can do very sophisticated and very accurate path-finding flights.”

The new design also addresses a common issue in robotic insects, i.e. lift efficiency. By positioning the wings to avoid interference from one another, the researchers have managed to maximise their lift force, thereby allowing for faster and more stable flight.

Why Is This Development Important?

The implications of these advancements could be far-reaching. Artificial pollination could become a practical solution in vertical farming, a growing industry focused on producing food in stacked indoor environments. For example, produce such as leafy greens (like lettuce and spinach), herbs (such as basil and mint), strawberries, tomatoes, and microgreens are commonly grown in vertical farms. These high-tech farms, often located in urban areas, aim to reduce agriculture’s environmental footprint by using less land and water while eliminating the need for chemical pesticides.

As the researchers point out: “Farmers in the future could grow fruits and vegetables inside multilevel warehouses, boosting yields while mitigating some of agriculture’s harmful impacts on the environment.” Robotic pollinators may also help maintain some crop yields in areas where natural pollinators are scarce or absent (albeit on a much smaller scale than our natural pollinators).

Beyond agriculture, and perhaps more realistically, the robots could be used in tasks such as inspecting hard-to-reach areas in machinery or infrastructure. Their ability to navigate tight spaces and perform precise movements makes them ideal for jobs that are hazardous or impossible for humans.

Limitations and Challenges

While the robots’ capabilities are impressive, there are significant hurdles to overcome before they can be deployed outside the laboratory. Currently, the robots rely on external power sources and control systems, as their size makes it difficult to integrate onboard batteries and sensors. Miniaturising these components remains a priority for Chen’s team, who aim to create fully autonomous flying robots within the next three to five years.

Another challenge lies in replicating the sophisticated muscle control of real insects. Bees, for example, can adjust their wing movements with incredible precision, allowing them to navigate complex environments with ease. While the MIT robots have made strides in this area, they still fall short of matching the natural agility and adaptability of their biological counterparts.

The introduction of robotic pollinators raises significant ethical and environmental questions. Critics caution that prioritising the development of these artificial systems risks diverting attention and resources from safeguarding the intricate, incredible network of natural pollinators that already exists. This amazing and incredible system, composed of bees, butterflies, birds, and countless other species, functions seamlessly on a global scale, providing pollination services that are sustainable, efficient, and free. Attempting to replicate such a complex and self-sustaining mechanism with robots not only seems far-fetched but also highlights the irreplaceable value of the natural world. Instead of relying on technological substitutes, there is a growing call to double down on efforts to restore and maintain the habitats and populations of these vital creatures, ensuring the resilience of ecosystems and food systems for generations to come.

Also, even if these robots could conceivably be produced at scale, widespread deployment of robotic insects could have unforeseen ecological consequences, e.g., disrupting existing ecosystems or creating new dependencies on artificial technologies.

The Road Ahead

Despite these challenges, the potential benefits of robotic pollinators are undeniable. The MIT team is already planning the next phase of development, which includes extending flight durations to over 10,000 seconds and improving the robots’ ability to land and take off from flowers. They are also exploring ways to incorporate sensors and computing capabilities, which would enable the robots to navigate and operate autonomously in outdoor environments.

“This new robot platform is a major result from our group and leads to many exciting directions,” says Chen. “For example, incorporating sensors, batteries, and computing capabilities on this robot will be a central focus in the next three to five years.”

A New Frontier in Sustainable Agriculture?

As the world grapples with the twin challenges of feeding a growing population and preserving biodiversity, innovations like MIT’s robotic pollinators offer a glimpse of a more sustainable future. While they are unlikely to replace natural pollinators entirely, these tiny flying machines could play a crucial supporting role in modern agriculture, particularly in controlled environments like vertical farms.

For now, the dream of swarms of robotic insects buzzing through greenhouses and fields remains just that, i.e. a dream. But with continued research and development, these miniature marvels could soon become an integral part of the agricultural landscape, helping to secure food supplies while reducing environmental impact.

What Does This Mean For Your Organisation?

The development of insect-inspired robotic pollinators by MIT is undeniably a remarkable feat of engineering and a testament to human ingenuity. These tiny flying machines demonstrate the power of technology to address some of the challenges posed by a changing world, particularly the growing threats to natural pollinator populations. With their improved agility, durability, and precision, these robots could open up possibilities for innovation in agriculture, infrastructure inspection, and beyond. However, their role as a potential substitute for nature’s intricate systems invites both excitement and caution.

While the robots could potentially aid in controlled environments like vertical farms or in regions where pollinator populations are critically low, it is crucial to acknowledge their limitations. At present, these robots remain highly experimental, reliant on external power sources and laboratory settings. Even with future advancements, the idea of deploying robotic swarms as a comprehensive replacement for the natural pollination system remains, at best, an extraordinary technical and ecological challenge. Natural pollinators, such as bees and butterflies, represent an intricate balance of biological and environmental systems that has evolved over millennia. Their efficiency, adaptability, and symbiotic relationship with ecosystems are unmatched by any human-made device.

Also, the ethical and environmental implications of relying on robotic pollinators cannot be ignored. For example, opting for technological solutions risks sidelining critical efforts to restore and preserve natural habitats, which are vital not only for pollinators but for the biodiversity and ecosystems that underpin life on Earth. Investing in the conservation of bees, butterflies, and other pollinating species is not merely an ecological responsibility but a pragmatic strategy to ensure the sustainability of agriculture and food production for the long term.

This is not to say that robotic pollinators lack value. Their potential to complement natural systems, most likely in niche or controlled environments, could prove invaluable. For example, in vertical farming, where natural pollinators cannot operate, these robots could contribute to sustainable urban agriculture. Similarly, their ability to perform precise, controlled manoeuvres in hazardous or inaccessible locations might unlock applications beyond pollination, such as infrastructure inspection and disaster response.

However, the broader focus should remain on addressing the root causes of pollinator decline, i.e., pesticide usage, habitat destruction, and climate change. These systemic issues require global collaboration, robust policy frameworks, and widespread public engagement. The preservation of natural pollinators and their habitats should be a central pillar of sustainability efforts, with technology serving as a complementary tool rather than a wholesale replacement.

The advancements in robotic pollinators are a powerful demonstration of human creativity and problem-solving. They offer promising opportunities in specific scenarios, but they should not distract from the urgent need to protect and restore the ecosystems that sustain natural pollinators. By balancing innovation with conservation, we can work towards a future where technology supports, rather than substitutes, the natural processes that are essential to life on Earth.

Which is better : DeepSeek or ChatGPT? In this video, Jonathan puts both through their paces and you can see the results for yourself.

[Note – To Watch This Video without glitches/interruptions, It may be best to download it first]

Use Edge’s built-in translation feature to quickly translate text into English directly in the web page you’re looking at, without having to copy it to another tool. Here’s how:

Highlight Text on a Web Page:

– Select the text you want to translate.

Use the Translation Option:

– Right-click the highlighted text and select Translate selection.

This saves time during research or communication with international colleagues.